In order to design and safely construct buildings and structures that are exposed to environmental processes such as earthquakes and wind loads, simulations are essential in advance. Although simulations are an approximation of reality, they are still dominated by uncertainties, that must be taken into account. These uncertainties can arise for various reasons, such as incorrectly recorded data or inaccurate simulation models. One widely used approach for generating and simulating environmental processes is the power spectral density (PSD) function. It establishes a relationship between the time and frequency domains and determines the relevant frequencies and their magnitude of the transformed signals. Since the model of the PSD function provides discrete values of the density for each frequency, the idea arises to model this density uncertain. For this purpose, statistical values are extracted from an ensemble of similar